Network Traffic Classification Based on Error-Correcting Output Codes and NN Ensemble

Xiao Xie, Bo Yang, Yuehui Chen, Lin Wang, Zhenxiang Chen

Abstract: Classification of network traffic is basic and essential for many network researches and managements. However, classification of network traffic using port-based and simple payload-based methods is diminished with the rapid development of peer-to-peer (P2P) application using dynamic port, disguising techniques and encryption to avoid detection. An alternative method based on statistics and machine learning has attracted researchers' attention in recent years. In this paper, a new approach based on the implementation of artificial neural network ensemble with the error-correcting output codes (ECOC) is proposed for classification of multi-class network traffic. As the error-correcting output codes have error correcting ability and improve the generalization ability of the base classifiers the experiments show that the proposed method can improve the multi-class classification accuracy by 12%-20% on datasets captured on the backbone router of our campus through a week.